U.S National Institutes fo Health | cancer.gov NCI Logo

Cancer Intervention and Surveillance Modeling Network

Modeling to guide public health research and priorities

Other Achievements: Highlights

What are the costs and benefits of screening mammography in younger and older women?

Several models have addressed the important question of which ages are best for starting and stopping mammography screening. The Dana-Farber Cancer Institute modeled the mortality benefits of screening women aged 40-49 and women over age 65, finding relatively small mortality benefits associated with screening younger women and larger mortality benefits for women over 65. Dana-Farber also examined the potential mortality benefits associated with periodic and staggered screening schedules (Lee 2004). The SPECTRUM model, developed at Georgetown University, estimated the costs and benefits of screening women from age 50 until age 70, 79, or for a lifetime. This analysis concluded that even under idealized treatment, the benefit of mammography after age 79 is too low relative to its cost to justify continued screening (Mandelblatt 2005).

Return to Top

What are the costs and benefits of interventions to improve breast cancer outcomes in African-American women?

Historically, African-American women have experienced higher breast cancer mortality rates than white women despite having lower incidence rates. The SPECTRUM model evaluated the costs associated with interventions intended to increase screening rates and promote the application of recommended treatment for African-American women. Outcomes included the number of mammograms, stage distribution at diagnosis, all-cause mortality, and discounted life years saved. Rather than investing in additional efforts to increase screening rates, ensuring that African-American women receive intensive treatment appears to be the most cost-effective approach to decreasing the disproportionate mortality rates (Mandelblatt 2004).

Return to Top

Looking back at mammography screening between 1990 and 2000—could we have done better?

The model developed at University of Wisconsin-Madison was used to perform a retrospective cost-effectiveness analysis of screening as it actually occurred in the 1990s as compared to how screening could have been performed under various strategies. Outcomes considered were costs and quality-adjusted life years associated with actual and hypothetical screening scenarios. The analysis concluded that less frequent mammograms in conjunction with more women participating in screening programs would result in more life years saved at a lower cost (Stout 2006).

Return to Top

Is it cost-effective to screen BRCA1/2 mutation carriers with breast magnetic resonance imaging (MRI)?

Women with an inherited BRCA1/2 mutation are at high risk for breast cancer, which mammography often misses. The model developed at Stanford University was used to estimate the survival benefit, incremental costs, and cost-effectiveness of MRI plus mammography screening strategies for BRCA 1/2 mutation carriers as compared to mammography alone. The analysis concluded that the mammography plus MRI screening is more cost-effective than mammography alone for BRCA1 and BRCA2 mutation carriers. The cost-effectiveness varied greatly by age, with greater benefits seen in younger women and women with dense breasts (Plevritis, Kurian et al. 2006).

Return to Top

Can the probability of overdiagnosis in early detection programs be estimated?

The probability of overdiagnosis, defined as the situation where a screening exam detects a disease that would have otherwise been undetected in a person's lifetime, is an ongoing concern with early detection of breast cancer. Researchers at Dana-Farber worked to derive the mathematical expression for the probability of overdiagnosis. Their methods have previously been used to estimate overdiagnosis for prostate cancer and are currently being applied to breast cancer (Davidov 2004). Since overdiagnosis can never be directly observed, indirect mathematical approaches are a key to providing estimates of this critical measure.

Return to Top